Hi, On 10/16/05, Zuxy <zuxy.meng@gmail.com> wrote:
Hi,
cputable.h.diff contains information update, bug fixes and punctuation modifications to the original cputable.h. I suppose there's no reason to reject it:-)
cpudetect.h.diff and cpudetect.c.diff together change the original behavior of mplayer: it first determines whether the CPU has a built-in namestring; if so, then the namestring is displayed and codenames in cputable.h aren't used; if not it'll fall back to the original behavior, looking up the cputable.h.
I've explained the reasons for the last two patches in another thread. However, I don't go radical by removing cputable.h at all and depending on built-in namestring only, partly because only P4 or newer Intel CPUs have a namestring (all AMD CPUs supporting cpuid support namestring), and in memory of those who have collected all the odd codenames as well:-)
The team are to decide whether to apply only cputable.h.diff, or apply all the three. In the latter case mplayer may contain dead codes and data, those that should never be executed or accessed on a commercially available CPU.
I'd much prefer to rely on the built-in namestring whenever it's available. In my case, here's how the old: CPUDETECT: CPU: Advanced Micro Devices (Family: 15, Stepping: 0) and new: CPUDETECT: CPU: AMD Athlon(tm) 64 Processor 3400+ (Family: F, Model: C, Stepping: 0) CPU detection looks like. The new form is much-much better IMHO. The code looks Okay to me. Comments? Guillaume -- Reading doesn't hurt, really! -- Dominik 'Rathann' Mierzejewski